© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 54824
    Prev
    Next
    INDEX
    Explanations

    phrases that denote power dynamics and related concepts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Powered
    -0.35
    powered
    -0.35
     powering
    -0.34
     powered
    -0.34
    -powered
    -0.34
     Powered
    -0.33
     powers
    -0.32
     powerful
    -0.32
     Powers
    -0.31
     Powerful
    -0.31
    POSITIVE LOGITS
     money
    0.20
     influence
    0.19
     Money
    0.18
    money
    0.18
     Prest
    0.17
     attorney
    0.17
    (PC
    0.17
     bald
    0.17
     MONEY
    0.17
     authority
    0.17
    Activations Density 0.011%

    No Known Activations